existing mechanism uses a global queue for some buffers and the
vp->b_dirtyblkhd queue for others. This turns sequential writes into
randomly ordered writes to the server, affecting both read and write
performance. The existing mechanism also copes badly with hung
servers, tending to block accesses to other servers when all the iods
are waiting for a hung server.
The new mechanism uses a queue for each mount point. All asynchronous
i/o goes through this queue which preserves the ordering of requests.
A simple mechanism ensures that the iods are shared out fairly between
active mount points. This removes the sysctl variable vfs.nfs.dwrite
since the new queueing mechanism removes the old delayed write code
completely.
This should go into the 2.2 branch.
/*
* Structure defined by POSIX.4 to be like a timeval.
*/
struct timespec {
time_t ts_sec; /* seconds */
long ts_nsec; /* and nanoseconds */
};
The correct names of the fields are tv_sec and tv_nsec.
Reminded by: James Drobina <jdrobina@infinet.com>
The interface into the "VMIO" system has changed to be more consistant
and robust. Essentially, it is now no longer necessary to call vn_open
to get merged VM/Buffer cache operation, and exceptional conditions
such as merged operation of VBLK devices is simpler and more correct.
This code corrects a potentially large set of problems including the
problems with ktrace output and loaded systems, file create/deletes,
etc.
Most of the changes to NFS are cosmetic and name changes, eliminating
a layer of subroutine calls. The direct calls to vput/vrele have
been re-instituted for better cross platform compatibility.
Reviewed by: davidg
rick@snowhite.cis.uoguelph.ca:
1. Clear B_NEEDCOMMIT in nfs_write to make sure that dirty data is
correctly send to the server. If a buffer was dirtied when it was in
the B_DELWRI+B_NEEDCOMMIT state, the state of the buffer was left
unchanged and when the buffer was later cleaned, just a commit rpc was
made to the server to complete the previous write. Clearing
B_NEEDCOMMIT ensures that another write is made to the server.
2. If a server returned a server (for whatever reason) returned an
answer to a write RPC that implied that fewer bytes than requested
were written, bad things would happen.
3. The setattr operation passed on the atime in stead of the mtime to
the server. The fix is trivial.
4. XIDs always started at 0, but this caused some servers (older DEC
OSF/1 3.0 so I've been told) who had very long-lasting XID caches to
get confused if, after a reboot of a BSD client, RPCs came in with a
XID that had in the past been used before from that client. Patch is
to use the current time in seconds as a starting point for XIDs. The
patch below is not perfect, because it requires the root fs to be
mounted first. This is because of the check BSD systems do, comparing
FS time to system time.
Reviewed by: Bruce Evans, Terry Lambert.
Obtained from: frank@fwi.uva.nl (Frank van der Linden) via rick@snowhite.cis.uoguelph.ca
Poul mentioned that he thought this was some kind of timing problem, and
that started me thinking. After a little poking around, I found that
nfs_timer() was completely disabled when NFS_NOSERVER was #defined.
But after looking at nfs_timer(), it seemed like it was something
required by both the client and server code, and disabling it outright
just didn't seem to make any sense. Parts of it relate only to the
NFS server side code, so I disabled those, but I re-enabled the rest
of the function and made sure that it would be called from nfs_init()
(in nfs_subs.c).
With nfs_timer() re-enabled, everything seems to work again. The only
other changes I made were to #ifdef away some variable declarations
in the NFS_NOSERVER case so that gcc would stop complaining about
unused variables.
Reviewed by: phk
Submitted by: Bill Paul <wpaul@skynet.ctr.columbia.edu>
Speed up for vfs_bio -- addition of a routine bqrelse to greatly diminish
overhead for merged cache.
Efficiency improvement for vfs_cluster. It used to do alot of redundant
calls to cluster_rbuild.
Correct the ordering for vrele of .text and release of credentials.
Use the selective tlb update for 486/586/P6.
Numerous fixes to the size of objects allocated for files. Additionally,
fixes in the various pagers.
Fixes for proper positioning of vnode_pager_setsize in msdosfs and ext2fs.
Fixes in the swap pager for exhausted resources. The pageout code
will not as readily thrash.
Change the page queue flags (PG_ACTIVE, PG_INACTIVE, PG_FREE, PG_CACHE) into
page queue indices (PQ_ACTIVE, PQ_INACTIVE, PQ_FREE, PQ_CACHE),
thereby improving efficiency of several routines.
Eliminate even more unnecessary vm_page_protect operations.
Significantly speed up process forks.
Make vm_object_page_clean more efficient, thereby eliminating the pause
that happens every 30seconds.
Make sequential clustered writes B_ASYNC instead of B_DELWRI even in the
case of filesystems mounted async.
Fix a panic with busy pages when write clustering is done for non-VMIO
buffers.
proc or any VM system structure will have to be rebuilt!!!
Much needed overhaul of the VM system. Included in this first round of
changes:
1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
haspage, and sync operations are supported. The haspage interface now
provides information about clusterability. All pager routines now take
struct vm_object's instead of "pagers".
2) Improved data structures. In the previous paradigm, there is constant
confusion caused by pagers being both a data structure ("allocate a
pager") and a collection of routines. The idea of a pager structure has
escentially been eliminated. Objects now have types, and this type is
used to index the appropriate pager. In most cases, items in the pager
structure were duplicated in the object data structure and thus were
unnecessary. In the few cases that remained, a un_pager structure union
was created in the object to contain these items.
3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
be removed. For instance, vm_object_enter(), vm_object_lookup(),
vm_object_remove(), and the associated object hash list were some of the
things that were removed.
4) simple_lock's removed. Discussion with several people reveals that the
SMP locking primitives used in the VM system aren't likely the mechanism
that we'll be adopting. Even if it were, the locking that was in the code
was very inadequate and would have to be mostly re-done anyway. The
locking in a uni-processor kernel was a no-op but went a long way toward
making the code difficult to read and debug.
5) Places that attempted to kludge-up the fact that we don't have kernel
thread support have been fixed to reflect the reality that we are really
dealing with processes, not threads. The VM system didn't have complete
thread support, so the comments and mis-named routines were just wrong.
We now use tsleep and wakeup directly in the lock routines, for instance.
6) Where appropriate, the pagers have been improved, especially in the
pager_alloc routines. Most of the pager_allocs have been rewritten and
are now faster and easier to maintain.
7) The pagedaemon pageout clustering algorithm has been rewritten and
now tries harder to output an even number of pages before and after
the requested page. This is sort of the reverse of the ideal pagein
algorithm and should provide better overall performance.
8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
have been removed. Some other unnecessary casts have also been removed.
9) Some almost useless debugging code removed.
10) Terminology of shadow objects vs. backing objects straightened out.
The fact that the vm_object data structure escentially had this
backwards really confused things. The use of "shadow" and "backing
object" throughout the code is now internally consistent and correct
in the Mach terminology.
11) Several minor bug fixes, including one in the vm daemon that caused
0 RSS objects to not get purged as intended.
12) A "default pager" has now been created which cleans up the transition
of objects to the "swap" type. The previous checks throughout the code
for swp->pg_data != NULL were really ugly. This change also provides
the rudiments for future backing of "anonymous" memory by something
other than the swap pager (via the vnode pager, for example), and it
allows the decision about which of these pagers to use to be made
dynamically (although will need some additional decision code to do
this, of course).
13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
object" code has been removed. MAP_COPY was undocumented and non-
standard. It was furthermore broken in several ways which caused its
behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
continue to work correctly, but via the slightly different semantics
of MAP_PRIVATE.
14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
threads design can be worked around in other ways. Both #12 and #13
were done to simplify the code and improve readability and maintain-
ability. (As were most all of these changes)
TODO:
1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
this will reduce the vnode pager to a mere fraction of its current size.
2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
information provided by the new haspage pager interface. This will
substantially reduce the overhead by eliminating a large number of
VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
improved to provide both a "behind" and "ahead" indication of
contiguousness.
3) Implement the extended features of pager_haspage in swap_pager_haspage().
It currently just says 0 pages ahead/behind.
4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
via a much more general mechanism that could also be used for disk
striping of regular filesystems.
5) Do something to improve the architecture of vm_object_collapse(). The
fact that it makes calls into the swap pager and knows too much about
how the swap pager operates really bothers me. It also doesn't allow
for collapsing of non-swap pager objects ("unnamed" objects backed by
other pagers).
that call vnode_pager_alloc() so that a failure return can be dealt with.
This fixes a panic seen on NFS clients when a file being opened is deleted
on the server before the open completes.
2) Removed unnecessary vm_object_lookup()/pager_cache(object, TRUE) pairs
after vnode_pager_alloc() calls - the object is already guaranteed to be
persistent.
3) Removed some gratuitous casts.
The version 2 support has been tested (client+server) against FreeBSD-2.0,
IRIX 5.3 and FreeBSD-current (using a loopback mount). The version 2 support
is stable AFAIK.
The version 3 support has been tested with a loopback mount and minimally
against an IRIX 5.3 server. It needs more testing and may have problems.
I have patched amd to support the new variable length filehandles although
it will still only use version 2 of the protocol.
Before booting a kernel with these changes, nfs clients will need to at least
build and install /usr/sbin/mount_nfs. Servers will need to build and
install /usr/sbin/mountd.
NFS diskless support is untested.
Obtained from: Rick Macklem <rick@snowhite.cis.uoguelph.ca>
is an ambiguity in the NFS version 2 protocol.
VREG should be taken literally as a regular file. If a
server intents to return some type information differently
in the upper bits of the mode field (e.g. for sockets, or
FIFOs), NFSv2 mandates fa_type to be VNON. Anyway, we
leave the examination of the mode bits even in the VREG
case to avoid breakage for bogus servers, but we make sure
that there are actually type bits set in the upper part of
fa_mode (and failing that, trust the va_type field).
NFSv3 cleared the issue, and requires fa_mode to not
contain any type information (while also introduing sockets
and FIFOs for fa_type).
The fix has been tested against a variety of NFS servers.
It fixes problems with the ``Tropic'' NFS server for Windows,
while apparently not breaking anything.
Pointed-out by: scott@zorch.sf-bay.org (Scott Hazen Mueller)
handled correctly. This would manifest itself as "object deallocated too
many times" panics and perhaps other strange inconsistencies on NFS servers.
Reviewed by: me, of course
Submitted by: John Dyson
much higher filesystem I/O performance, and much better paging performance. It
represents the culmination of over 6 months of R&D.
The majority of the merged VM/cache work is by John Dyson.
The following highlights the most significant changes. Additionally, there are
(mostly minor) changes to the various filesystem modules (nfs, msdosfs, etc) to
support the new VM/buffer scheme.
vfs_bio.c:
Significant rewrite of most of vfs_bio to support the merged VM buffer cache
scheme. The scheme is almost fully compatible with the old filesystem
interface. Significant improvement in the number of opportunities for write
clustering.
vfs_cluster.c, vfs_subr.c
Upgrade and performance enhancements in vfs layer code to support merged
VM/buffer cache. Fixup of vfs_cluster to eliminate the bogus pagemove stuff.
vm_object.c:
Yet more improvements in the collapse code. Elimination of some windows that
can cause list corruption.
vm_pageout.c:
Fixed it, it really works better now. Somehow in 2.0, some "enhancements"
broke the code. This code has been reworked from the ground-up.
vm_fault.c, vm_page.c, pmap.c, vm_object.c
Support for small-block filesystems with merged VM/buffer cache scheme.
pmap.c vm_map.c
Dynamic kernel VM size, now we dont have to pre-allocate excessive numbers of
kernel PTs.
vm_glue.c
Much simpler and more effective swapping code. No more gratuitous swapping.
proc.h
Fixed the problem that the p_lock flag was not being cleared on a fork.
swap_pager.c, vnode_pager.c
Removal of old vfs_bio cruft to support the past pseudo-coherency. Now the
code doesn't need it anymore.
machdep.c
Changes to better support the parameter values for the merged VM/buffer cache
scheme.
machdep.c, kern_exec.c, vm_glue.c
Implemented a seperate submap for temporary exec string space and another one
to contain process upages. This eliminates all map fragmentation problems
that previously existed.
ffs_inode.c, ufs_inode.c, ufs_readwrite.c
Changes for merged VM/buffer cache. Add "bypass" support for sneaking in on
busy buffers.
Submitted by: John Dyson and David Greenman
- Make a number of filesystems work again when they are statically compiled
(blush)
- FIFOs are no longer optional; ``options FIFO'' removed from distributed
config files.
